OBJECTIVE: Using both general growth balance (GGB) and synthetic extinct generations (SEG) methods to evaluate the underreporting of deaths in disease surveillance points (DSP) from 1991 to 1998. METHODS: We used those two methods to estimate the underreport rate in DSP from 1991-1998. According to GGB method, death rate and the difference between entry rate and growth rate were regarded as independent and dependent variable, respectively, to fit a one-dimensional linear equation n(*)(x) - r(*)(x+) = [ln(k(1)/k(2))]/t + [(k(1)k(2))(0.5)/c]d(*)(x+). About SEG method, it was calculated by life expectancy at age 85 of national census in 2000 and an equation N(x) = integral(x)(infinity)D(a).exp[integral(x)(a)r(u)du]d(a). Then we drew scatter diagram to evaluate underreporting of average intercensal age-specific mortality rates, and made adjustment. RESULTS: By GGB method, the population count in 1998 was 87.29 percent as complete as the count in 1991. For the period of 1991 - 1998, the number of underreporting of deaths was 11.7 percent. The underreporting of average intercensal age-specific mortality rates was not high. However, when SEG method was used, the underreport rate became 48.0 percent. The underreporting of average intercensal age-specific mortality rates decreased significantly from younger to older age groups. CONCLUSION: GGB and SEG could reduce more cost than other underreporting-related survey, and evaluate the underreporting in a defined time period. GGB could also evaluate the underreporting rate of population and death. However, both these two methods had some limitations. There was big difference between the results when using SEG and GGB, suggesting that we should try to combine GGB and SEG methods to get the better results.
OBJECTIVE: Using both general growth balance (GGB) and synthetic extinct generations (SEG) methods to evaluate the underreporting of deaths in disease surveillance points (DSP) from 1991 to 1998. METHODS: We used those two methods to estimate the underreport rate in DSP from 1991-1998. According to GGB method, death rate and the difference between entry rate and growth rate were regarded as independent and dependent variable, respectively, to fit a one-dimensional linear equation n(*)(x) - r(*)(x+) = [ln(k(1)/k(2))]/t + [(k(1)k(2))(0.5)/c]d(*)(x+). About SEG method, it was calculated by life expectancy at age 85 of national census in 2000 and an equation N(x) = integral(x)(infinity)D(a).exp[integral(x)(a)r(u)du]d(a). Then we drew scatter diagram to evaluate underreporting of average intercensal age-specific mortality rates, and made adjustment. RESULTS: By GGB method, the population count in 1998 was 87.29 percent as complete as the count in 1991. For the period of 1991 - 1998, the number of underreporting of deaths was 11.7 percent. The underreporting of average intercensal age-specific mortality rates was not high. However, when SEG method was used, the underreport rate became 48.0 percent. The underreporting of average intercensal age-specific mortality rates decreased significantly from younger to older age groups. CONCLUSION: GGB and SEG could reduce more cost than other underreporting-related survey, and evaluate the underreporting in a defined time period. GGB could also evaluate the underreporting rate of population and death. However, both these two methods had some limitations. There was big difference between the results when using SEG and GGB, suggesting that we should try to combine GGB and SEG methods to get the better results.